Equipment monitoring support device, method, and program

Abstract

To perform processing for supporting monitoring of a facility when the facility operates while preventing complication and an increase in cost.SOLUTION: A facility monitoring support device 100 includes: an input unit 101 that receives an input of state information from sensors 1 and 2 which are installed at a facility and obtain plural kinds of state information; a determination unit 102 that determines whether the facility is operating or not operating, based on the state information that is input through the input unit 101; and a monitoring support unit 103 that performs, only when the determination unit 102 determines that the facility is operating, processing for supporting monitoring of the facility, based on the state information that is input through the input unit 101.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to an equipment monitoring support device, method, and program for supporting the monitoring of equipment. [Background technology]

[0002] As a technology related to equipment monitoring, Patent Document 1 discloses a technology for acquiring status information measured by sensors installed at multiple locations on a target part included in the equipment, and calculating an index a based on the acquired status information at multiple locations. PE It is disclosed that the indicator a is calculated. PE is expressed by, for example, equation (1), where i, j are the identifiers of the sensors, N (N≧2) is the number of sensor installation locations (= the number of sensors), and x i , x j is the state information measured by sensors i and j, and t is time. When the equipment is in operation and the target part is in a stable state, the index a PE The value of is relatively small and its fluctuation is small, but when the target part changes to an unstable state, the potential energy between the data becomes large, and the index a PE In this way, at each time t, the index a expressed by equation (1) PE By calculating the above, it is possible to monitor the change in the state of the target part.

[0003]

number

[0004] [Patent Document 1] Patent Publication No. 2021-189756 [Non-patent literature]

[0005] [Non-Patent Document 1] Sebastian Raschka: [Second Edition] Python Machine Learning Programming: Theory and Practice by a Master Data Scientist Summary of the Invention [Problem to be solved by the invention]

[0006] In equipment, there are operation modes such as operation and non-operation. Figure 4(a) shows that the index a PE When calculating the index a PE As shown in FIG. 4(a), when the operation mode is changed from operation to non-operation, the index a PE The change in the operating mode is equivalent to a change in the state of the target part, so the index a PE However, the change in the state of the target part when the operation mode changes does not necessarily mean that the condition is unstable. And, since changes in the operation mode occur on a daily basis, the index a PE It is desirable to avoid falsely detecting that the state is unstable.

[0007] As such, many of the technologies for monitoring equipment, including Patent Document 1, are originally intended to capture changes in the state of the equipment while it is in operation, and there is a demand for monitoring the equipment while the equipment is in operation. To achieve this, for example, a configuration is conceivable in which information on the equipment's operating state is acquired from an external host computer or the like to determine when the equipment is in operation, but this requires a mechanism such as building a network, which increases complexity and costs.

[0008] The present invention has been made in consideration of the above-mentioned points, and aims to enable processing to be performed to assist in monitoring of equipment while the equipment is in operation, while avoiding complexity and increased costs. [Means for solving the problem]

[0009] The facility monitoring support device of the present invention is installed in the facility, Multiple species From sensors that measure status information, Multiple species an input means for inputting the status information; Multiple species said status information only Based on A model that determines whether the equipment is in operation or not a determination means for determining whether the equipment is in operation or not, and when the determination means determines that the equipment is in operation, the input means Multiple species said status information Using at least one of and a monitoring support means for carrying out a process for supporting the monitoring of the facility. The facility monitoring support method of the present invention includes: Multiple species From sensors that measure status information, Multiple species an input step of inputting the state information; Multiple species said status information only Based on A model that determines whether the equipment is in operation or not a determination step of determining whether the equipment is in operation or not, and if it is determined in the determination step that the equipment is in operation, inputting the information input in the input step Multiple species said status information Using at least one of and a monitoring support step of carrying out a process for supporting the monitoring of the facility. The program of the present invention is installed in the facility, Multiple species From sensors that measure status information, Multiple species an input means for inputting the status information; Multiple species said status information only Based on A model that determines whether the equipment is in operation or not a determination means for determining whether the equipment is in operation or not, and when the determination means determines that the equipment is in operation, the input means Multiple species said status information Using at least one of The computer functions as a monitoring support means that performs processing to support the monitoring of the facility. [Effects of the Invention]

[0010] According to the present invention, it is possible to perform processing to support the monitoring of equipment while the equipment is in operation, while avoiding complexity and increased costs. [Brief explanation of the drawings]

[0011] [Figure 1] 1 is a diagram illustrating a functional configuration of an equipment monitoring support device according to an embodiment. [Figure 2] FIG. 2 is a diagram illustrating a schematic configuration of a tension reel. [Figure 3] FIG. 10 is a characteristic diagram showing an example of the distribution of status information when the tension reel is in operation and when it is not in operation. [Figure 4] FIG. 10 is a characteristic diagram showing an example of a time series change in the index aPE. DETAILED DESCRIPTION OF THE INVENTION

[0012] Hereinafter, preferred embodiments of the present invention will be described with reference to the accompanying drawings. FIG. 1 shows the functional configuration of an equipment monitoring support device 100 according to an embodiment. The equipment monitoring support device 100 includes an input unit 101 , a determination unit 102 , a monitoring support unit 103 , and an output unit 104 .

[0013] An input unit 101 receives input of status information from sensors 1 and 2 that are installed in the facility and measure the status information. In this embodiment, the monitoring support unit 103 uses the index a disclosed in Patent Document 1. PE Therefore, it is necessary to perform simultaneous multipoint measurements at multiple locations in the target part of the equipment, and multiple sensors 1 that measure predetermined state information (for example, temperature) are installed in the equipment. Furthermore, the determination unit 102 determines whether the equipment is operating or not based on multiple types of status information. Therefore, in addition to the sensor 1, at least one sensor 2 that measures other status information (for example, vibration) is installed in the equipment. The input unit 101 acquires, for example, at regular intervals, measured values ​​of the state information from the sensors 1 and 2. Note that communication between the sensors 1 and 2 and the equipment monitoring support device 100 may be wireless communication or wired communication.

[0014] The determination unit 102 determines whether the equipment is operating or not based on the state information input by the input unit 101 . In this embodiment, the determination unit 102 determines whether the equipment is operating or not by logistic regression analysis based on multiple types of state information (e.g., temperature, vibration) measured by the sensors 1 and 2 and input by the input unit 101. Details of the determination process by the determination unit 102 will be described later.

[0015] The monitoring support unit 103 performs processing to support the monitoring of the equipment based on the status information input by the input unit 101 only when the determining unit 102 determines that the equipment is in operation. In this embodiment, only when the determining unit 102 determines that the equipment is in operation, the monitoring support unit 103 determines whether the equipment is in operation, based on the state information (for example, temperature) input by the input unit 101 and measured at multiple points simultaneously by the multiple sensors 1, and calculates the index a disclosed in Patent Document 1. PE Calculate. In addition, in the monitoring support unit 103, as described in Patent Document 1, the index a PE It is also possible to calculate only the index a PE It is also possible to calculate the index a and compare it with a threshold value. PE Instead, a statistical analysis indicator SD Furthermore, the processing performed by the monitoring support unit 103 is not limited to the processing disclosed in Patent Document 1, as long as it supports the monitoring of the equipment based on the status information input by the input unit 101.

[0016] The output unit 104 outputs the result of the monitoring support unit 103. For example, an indicator a PE In addition, the result of the determination by the determination unit 102 may be output.

[0017] The determination process in the determination unit 102 will be described below. Here, an example will be described that targets a tension reel 200 used in a steel plate processing line, as shown in Figure 2. In a steel plate processing line, it is necessary to apply tension to both ends of a steel plate 204 to unwind and take it up, and a winding device called a tension reel 200 is used for winding. In the tension reel 200, a motor 201 is used as a drive source to drive a winding drum 203 via a reducer 202, thereby winding up the steel plate 204.

[0018] A sensor 1 that measures temperature and a sensor 2 that measures vibration are installed at appropriate locations on the motor 201 of the tension reel 200. Although sensors 1 are installed in multiple locations, only one sensor 1 is shown here in order to explain the process of determining whether the equipment is operating or not.

[0019] FIG. 3 shows an example of the distribution of status information when the tension reel 200 is in operation and when it is not in operation. FIG. 3(a) is a characteristic diagram showing the probability density of temperature data measured by sensor 1, and (b) is a characteristic diagram showing the probability density of vibration data measured by sensor 2. As shown in FIGS. 3(a) and (b), it was confirmed that both the temperature data and the vibration data behave differently when in operation and when not in operation. Specifically, as shown in FIG. 3(a), the temperature when not in operation is generally concentrated in the range of 10 to 15°C, while the temperature when in operation is generally concentrated in the range of 20 to 30°C. Furthermore, as shown in FIG. 3(b), the vibration when not in operation is generally 0 m / s 2 vibrations during operation are generally 2 to 10 m / s 2 are concentrated in the range of

[0020] In this way, a model is created to determine whether a plant is operating or not based on the changes in temperature data and vibration data that occur depending on the operating mode, such as operating or not. In this classification problem, the objective variable is binary qualitative data, and the explanatory variable is quantitative data. The operating mode is estimated by utilizing logistic regression analysis, which is known as a method for solving such classification problems. Specifically, the operating / non-operating status is determined by equation (2): β0 to β2 are regression coefficients, π(t) is the probability that the objective variable is "operating", x1 and x2 are the measured state information, namely, temperature data and vibration data, respectively, and t is the time.

[0021]

number

[0022] The results of the operation mode estimation using logistic regression analysis showed a good accuracy, with an operation accuracy rate of 99.9% and a correct answer rate of 96.0%. The correct answer rate is the rate at which the classification results for all data were correct. The operation accuracy rate is the probability that the data that the model classified as operating was actually operating.

[0023] It is also possible to use decision tree analysis or gradient boosting instead of logistic regression analysis. The results of the operation mode estimation using decision tree analysis showed an operation precision rate of 97.2% and an accuracy rate of 95.2%. Furthermore, the results of the operation mode estimation using gradient boosting showed an operation precision rate of 96.4% and an accuracy rate of 95.8%. As described above, logistic regression analysis, decision tree analysis, and gradient boosting all achieved good accuracy, with both the operational precision rate and accuracy rate exceeding 90%. Note that logistic regression analysis, decision tree analysis, and gradient boosting are well-known technologies, and detailed explanations thereof will be omitted (see, for example, Non-Patent Document 1).

[0024] FIG. 4(b) shows the index a calculated by the monitoring support unit 103. PE As shown in FIG. 4(b), the index a is set to 0 only when the determining unit 102 determines that the equipment is in operation. PE In other words, when the determining unit 102 determines that the equipment is not in operation, the index a PE As a result, as mentioned in Figure 4(a), whenever the operation mode changes, the index a PEThis can avoid false detection that the state is unstable.

[0025] As shown in Figure 4(a), when the operation mode is not taken into account, and as shown in Figure 4(b), when the operation mode is taken into account, the target part of the equipment is not in an unstable state, but the index a PE The false alarm rate exceeded the specified threshold. As a result, the false alarm rate when the operation mode was not taken into account was 10.30%, while the false alarm rate when the operation mode was taken into account was 1.12%. In this way, by taking the operation mode into account, the index a was set only when it was determined that the equipment was in operation. PE By calculating the above, false detection can be suppressed.

[0026] As described above, the operation or non-operation of the equipment is determined based on the measured values ​​of the status information from the sensors 1 and 2, so that there is no need for a mechanism such as building a network with an external host computer, etc., and it is possible to avoid complication and increase in cost. Then, the indicator a is set only when the determining unit 102 determines that the equipment is in operation. PE In other words, it becomes possible to perform processing to support monitoring of the equipment when the equipment is in operation.

[0027] Note that the index a is set only when the determining unit 102 determines that the equipment is in operation. PE However, the present invention is not limited to this. For example, the index a is calculated both when the determining unit 102 determines that the equipment is in operation and when the determining unit 102 determines that the equipment is in non-operation. PE is calculated, but the indicator a PE When the equipment is not in operation, PE It may also be possible to present the information in a way that makes it clear what the purpose of the information is.

[0028] An equipment monitoring support device to which the present invention is applied is configured by a computer device equipped with, for example, a CPU, ROM, RAM, etc., and the functions of each means are realized by the CPU executing a predetermined program stored, for example, in the ROM. The present invention can also be realized by supplying software (program) that realizes the functions of the present invention to a system or device via a network or various storage media, and having the computer of that system or device read and execute the program. [Explanation of symbols]

[0029] 100: Equipment monitoring support device, 101: Input unit, 102: Determination unit, 103: Monitoring support unit, 104: Output unit

Claims

1. an input means for inputting a plurality of types of status information from sensors installed in the facility that measure the plurality of types of status information; a determination means for determining whether the equipment is in operation or not using a model for determining whether the equipment is in operation or not based only on the plurality of types of status information input by the input means; and a monitoring support means for, when the determination means determines that the equipment is in operation, using at least one of the multiple types of status information input by the input means, carrying out processing to support monitoring of the equipment.

2. The equipment monitoring support device described in Claim 1, characterized in that the determination means determines whether the equipment is operating or not by logistic regression analysis, decision tree analysis, or gradient boosting.

3. 3. The equipment monitoring support device according to claim 1, wherein the monitoring support means performs processing to support monitoring of the equipment using at least one of the multiple types of status information input by the input means only when the determination means determines that the equipment is in operation.

4. an input step of inputting a plurality of types of status information from sensors installed in the equipment that measure the plurality of types of status information; a determination step of determining whether the equipment is in operation or not using a model that determines whether the equipment is in operation or not based only on the multiple types of status information input in the input step; and a monitoring support step of, when it is determined in the determination step that the equipment is in operation, carrying out processing to support monitoring of the equipment using at least one of the multiple types of status information input in the input step.

5. an input means for inputting a plurality of types of status information from sensors installed in the facility that measure the plurality of types of status information; a determination means for determining whether the equipment is in operation or not using a model for determining whether the equipment is in operation or not based only on the plurality of types of status information input by the input means; A program for causing a computer to function as monitoring support means that, when the determination means determines that the equipment is in operation, performs processing to support monitoring of the equipment using at least one of the multiple types of status information input by the input means.

Images